Author: Robert Hyatt
Date: 06:26:19 06/28/99
Go up one level in this thread
On June 28, 1999 at 08:29:36, Francesco Di Tolla wrote: >>I can only offer two data points. In Crafty, _my_ timing allocation code >>assumes that there will actually be more time to use than crafty has at the >>point it has to make the decision. > >I'm sorry Bob, but I don't agree here. I think that if your program does this it >is a "problem" of your program. Any program that supports a "ponder off" mode >should also be able to play at maximum efficency within the time given. > Sorry but you are wrong. Should I spend 1/2 my time testing with ponder=off to be sure that it works optimally? Or should I spend 100% of my time testing in the way where the program plays the strongest? ponder=off was done to allow me to do some testing on one machine with two programs, but for _debugging_, not anything else. If I thought that this mode would be used to play 'real games' I would either test more (wasted time) or eliminate ponder=off completely. _either_ of those options would be bad IMHO... >Of course if you have more time (while waiting for opponents move) you can play >better, but in my personal opinion, if you assume that time is always there and >speculate on this it is your fault. > >(This has nothing to do with trusting eng-eng matches, but is a general >consideration). > >[...] > >>The other data point was an old Rebel. Ed did all of his time setting while >>"permanent brain" was being used. When he turned it off in the NPS match he >>saw bad time allocation too. > >again, something I would call "a bug" not a feature, or do newer version behave >the same? > >[...] > >>Another point is that some programs (again, mine is an example) depends on >>fairly fast hardware, because of some of the search decisions I have made in >>the design process. [...] > >This is reasonable: tweaking a program for the hardware you run on is correct... > >>Your "ponder=off" type matches effectively cut the processor speed by 1/2, >>when you think about it. And the results can definitely be affected.. > >... but this is not true, because you're looking at a different position, and >even if you make assumptions on next move, you'll have a percentage of correctly >guessed moves which is less than 100%, so the speed difference is less then one >half. > >Moreover, is it better to compare two engines on two PC's with a P200 MMX (like >SSDF fellows do) or let them coexist on a PII 400? >Assuming this two cpu have a factor 2 of difference, would it be better to spend >3 minutes on a move and 3 on the move of your opponent, or six on yours and none >on the other? > >Finally, do the SSDF guys compare properly progrmas making them play on Intel's >Pentiums when some engines (like Rebel) prefer other (AMD...). > >I have no real answer, so I think the important thing is not to take this >comparisons to seriously, especially to don't buy after the positioning in the >SSDF list! > >(Personally I choose the engine/program which gives me most fun when I play with >it.) > >regards >Franz
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.